Editors’ note: Today’s post is by Akhilesh Ayer, CEO of Cactus Technology Solutions (CACTUS), formerly known as Cactus Communications.

Traditionally, research integrity was perceived as a series of isolated failures. However, as the numbers suggest, this is changing. In 2002, approximately 1 in 5,000 published papers were retracted, but this number jumped to 1 in 500 by 2023. In fact, in 2023 alone, the number of retractions crossed 10,000, setting a new record. Retractions, however, only tell us what has already slipped through the cracks. So, for every withdrawn paper, there may be several that remain undetected. At the same time, an enormous number of questionable papers are being produced. Between 2020 and 2026, paper mills posted more than 18,700 advertisements selling authorship of research papers.

Taken together, these are not isolated warning signs. They raise some crucial questions: Should we continue to treat integrity failure as a problem of detection where we identify offenders, remove the bad papers, and tighten the rules? Or is it time to recognize that research integrity challenges are a result of the competition, scale, and complexity of modern research? And if the problem is systemic, who should address it?

The Growing Scale of Integrity Challenges

Misconduct is no longer episodic. It is becoming easier to produce papers of questionable quality at scale. As a result, researchers, publishers, funders, and institutions are facing challenges on different fronts. Publishers are under pressure to process huge volumes of submissions. They are also responsible for screening out questionable research from getting published. Editorial systems are strained with the responsibility of identifying high-quality research. As a result, peer review systems have slowed down, and publication delays have become more common. Slowly, over time, this is influencing researcher behavior, institutional incentives, and even funding decisions.

To address these problems, editorial workflows have evolved to identify sophisticated forms of misconduct. These include AI-generated text, manipulated images, and citation cartels, among other forms of research manipulation. These interventions are an important form of defense to preserve research integrity and identify complex forms of misconduct. However, they are ultimately designed to identify problems that have already entered the system. The long-term solutions would require addressing the conditions that allow problematic papers to be produced and submitted at scale.

The Economics of Academic Pressure

At the center of this shift is the familiar tension where the system demands rigor but rewards output. The “publish or perish” culture continues to define career progression, which creates immense pressure on researchers. Paper mills and ghostwriting services have leveraged this gap. They offer faster routes to publication and guarantee authorship. As a result, what requires effort and intent can be outsourced with ease and scaled up as needed. In fact, paper mills have achieved a level of sophistication where they also offer patents and placements in textbooks. At the same time, AI tools have further amplified the problem. Some AI tools make it possible to generate authentic looking papers in seconds. The result is a growing volume of work that looks legitimate and passes initial scrutiny even though the foundations are questionable.

The speed of output, the volume of publications, and journal reputation are considered the currency of academic careers. This is a misalignment that academia acknowledges but has found challenging to address. To resolve this at a system level, it is important for institutions to focus on how the researcher has advanced knowledge instead of how many papers they have produced. Promotions and funding are dependent on the volume. Instead, could researchers highlight some of their most important contributions over the past few years to be recognized for the significance of their work rather than for volume? If the incentive model changes, there is hope that the pressure will ease off at least a little bit.

Additionally, can publishers, institutions, universities, and research integrity bodies come together to take action against organized forms of research misconduct? Instead of focusing on individual researchers who participate in questionable practices, can the research community relook the incentives and systems that make it possible for misconduct to operate at scale?

The Cost of Unreliable Science

It is also important to consider the cost of unreliable science. Research misconduct today is the byproduct of a system where pressure, complexity, and scale make questionable shortcuts accessible and, in many ways, more rational. The result is a growing lack of trust in the system. The coexistence of flawed and reliable work that is largely indistinguishable can have a far-reaching impact. When the underlying work is uncertain, citations start to lose meaning. Even journals aren’t immune to the consequences of misconduct. Their prestige diminishes if publications aren’t rigorous. Beyond academia, policy frameworks and public trust in science depend on the assumption that published research is credible.

Fixing the system goes well beyond looking for better detection tools. It needs a shift in how responsibility is understood across the research lifecycle. Most current interventions are downstream where publishers play an important role, but opportunities exist to address issues at their source.

Moving from Detection to Prevention

Can universities play a more active role in addressing quality-related issues? Instead of relying on journals to identify problems, can they implement pre-submission quality review processes? Institutions could move toward conducting reviews for publication readiness, whereby manuscripts undergo internal checks before entering the publication process. Such an approach would shift responsibility upstream and strengthen the institution’s role in ensuring research quality.

This approach can be supported by emerging technology. AI-assisted screening tools can help institutions identify potential issues in manuscripts early, which would allow researchers to address them before submitting to journals. Such tools can complement human oversight and create an additional layer of quality assurance. However, technology alone may not be enough. Guiding researchers at every stage of writing and publication can help them uphold high standards of integrity.

To make improvements in research integrity sustainable, screening tools and checks may not be enough. It is important to incentivize responsible researcher behavior too. Today, researchers strive to build a strong publication record. However, with the growing concerns around research misconduct, it is worth considering whether researchers should build an “integrity record” or “integrity passport.” Such a record would demonstrate a researcher’s commitment to following responsible practices. Further, this would support them with career progression and funding applications.

Strengthening Integrity in Publishing

Further along the lifecycle, when manuscripts enter the publication system, publishers face the challenge of maintaining trust and credibility at scale. They need to face the difficult task of balancing efficiency while ensuring credibility. The growing volume of submissions, paired with limited visibility into the origins of manuscripts, poses a real challenge. A combination of editorial judgment and advanced integrity screening technologies would result in a more reliable screening process. Some of the latest screening tools can identify image manipulation, unusual citation patterns, ethical compliance issues, author identity concerns, and other indicators of research reliability. Such thorough checks can prevent problematic papers from entering the system. Beyond this, dealing with industrialized misconduct may require greater collaboration among publishers. A shared infrastructure that detects known indicators of misconduct such as citation manipulation patterns and fabricated identities could create a stronger defense against research misconduct.

Usually, it is high-profile scandals or mass retractions that raise eyebrows and spark concerns. However, real dents to research integrity are much quieter. The reality is that for the system to become ineffective, it does not need to collapse; it only needs to become unreliable. The challenge doesn’t end with identifying problematic papers. The underlying systems need to be strengthened, too. Publishers play an important role in keeping misconduct at bay, but integrity issues threaten the entire ecosystem. Therefore, all stakeholders in academia, including institutions, publishers, funders, and technology providers, need to put in a coordinated effort to address this challenge. It is this collective responsibility that will restore and protect the credibility of research.

Author’s note (AI use disclosure): AI tools were used for outlining the piece, proofreading, and language refinement. The ideas, arguments, source selection, and final decisions are the author’s own.